⭐ Available in the following bundle: Insights Bundle
⭐ Bundle included in the following plans: Scale
⭐ Available on the following legacy plans: Agency, Business
👀 Not sure which plan you're on? Check your subscription here.
A/B Testing is a built-in Heyflow feature that helps you optimize your funnel's performance by comparing different variants.
You can fully design both versions, run them against each other, review detailed statistics, and make data-driven decisions based on the better-performing version.
Check out our FAQ about A/B tests or learn how to run an A/B test in six easy steps below (click on each step to get more details👇):
How to run an A/B test with your flow
1. Build your heyflow
Before you can start an A/B test, you need to create your baseline flow — the original version that you'll be comparing the variant against. Build your flow as usual, ensuring that the baseline represents your current design and content.
2. Create A/B test and prepare the variant
Once your baseline flow is ready, click Create A/B Test to begin the setup.
Now it's time to create a variant of your baseline flow. You can modify anything on any page — add or delete blocks and pages, change the design or use a different copy. If needed, you can also make changes to the baseline flow itself. Make sure to switch to the right variant when editing your flow!
💡 Tip: While experimenting is valuable, it's important not to make overly drastic design differences between the baseline and the variant to get comparable results.
3. Launch the A/B test and choose your traffic distribution
When you’re done building and designing your flow variants, click on “Launch A/B test” in the top right corner. This will publish both variants of your flow.
During this step, you can adjust the traffic distribution to decide how much traffic should go to the baseline flow and the variant flow. By default, the traffic is split 50/50. You can always update the traffic distribution later in the Analytics Dashboard if needed.
❗ Important: The different variants are assigned using technical cookies. Consider adding a cookie consent to your flow to comply with legal and privacy requirements. You can find more information on Cookie Consent here.
💡 After you click Create A/B Test, your test is saved as a draft. Your live flow keeps running unchanged and no traffic goes to the variant until you click Launch A/B test, so you can prepare the variant over several sessions. While the test is still a draft you can discard it via the three-dot menu in the Launch A/B test dropdown. Once launched, a test can only be ended with Finish A/B test.
You can also unpublish your flow without removing the test.
4. Monitor performance analytics
While the A/B test is running, open Results. You land on the A/B Tests page, where you can follow visits, submits, conversion rate and the probability of being best for both versions in real time. The Analytics Dashboard keeps showing the combined data of your flow.
🔎 Learn more about A/B Testing and Analytics here.
5. End test and proceed with the better-performing version
Once you have enough data, click Finish A/B test in the top right corner and choose the version you want to keep. That version becomes your live flow and the other one is discarded. Your test results stay available on the A/B Tests page.
FAQ - Frequently Asked Questions
What is an A/B test?
An A/B test is a controlled experiment where two versions of a flow are directly compared to determine which one performs better in terms of user engagement and conversions. It helps you make data-driven decisions about your design and content choices.
Why should I run one/ what are the advantages of an A/B test?
Running an A/B test allows you to optimize your flow's effectiveness. It helps you understand what design and content changes resonate with your audience, leading to higher conversion rates and improved user experiences. Or ultimately, a better return on your investment.
How does an A/B test work?
An A/B test splits your traffic between two versions of your flow: the baseline and the variant. While a 50/50 traffic split is common, you can customize the split to suit your needs.
On the A/B Tests page under Results, you can track and compare how users interact with each version. This helps you identify which design and content elements perform better, so you can optimize your flow for success.
🔎 Visitors are assigned to a version at random. With a 50/50 split the numbers will be close over time, but they are rarely exactly equal, and single days can differ a lot, especially with low traffic or campaign spikes. Compare conversion rates, not absolute visits.
What to consider before you get started
When creating variants, avoid making drastic design differences that could skew the results. Focus on testing specific elements to gain accurate insights.
What can I change about the variant?
You can change pretty much anything about your variant — add or delete blocks and pages, change design elements like fonts or colors, or use a different copy.
You have Heyflow’s full building flexibility when designing your flow variant.
How long should I run an A/B test?
The duration of an A/B test depends on factors like your flow's traffic volume and the changes you're testing. It's generally recommended to run tests long enough to gather statistically significant data, which might vary based on your goals and audience. We recommend running your A/B tests for at least one week, in order to have an evenly distributed user behavior.
What happens with the traffic data once the A/B test is completed?
All relevant data is stored after a test has ended. You can select tests from the last 3 months in the dropdown on the A/B Tests page.
I keep on reloading my flow but always get the same variant
Technical cookies select the different variants. To get a different variant, clear your cookies or try in a different browser.
Can I duplicate a flow with an active A/B test?
Yes, you can duplicate a flow with an active or created A/B test. The duplication process will work like a normal duplication and create a new flow based on the baseline of the duplicated flow. The duplicated flow will not feature an active or created A/B test.







